The impact of green innovation on carbon reduction efficiency in China: Evidence from machine learning validation

This study analyzes the environmental dynamics in the Yangtze River Economic Belt from 2006 to 2020, using panel data from 108 cities. Employing the Modified Undesirable Epsilon-based measure approach, it assesses pollution reduction and carbon efficiency through a spatial evolution analysis. Advanced models, including fixed-effects, moderation effects, and threshold effects models, explore the impact and mechanisms of green technological innovation. Machine learning methods and a biased effects model further investigate the dynamic impact of green technology innovation. Key findings indicate that green technological innovation significantly enhances pollution reduction and carbon efficiency, especially in middle reaches, low-carbon, and non-resource cities. Formal and informal environmental regulations act as substantial moderators with varying efficacy. A single threshold effect based on development levels highlights varied moderating influences. Optimal factor input points are identified for green technology innovation, formal environmental regulation, and informal environmental regulation. Policy recommendations emphasize the need to enhance green technological innovation and implement tailored environmental regulatory frameworks to boost pollution reduction and carbon efficiency in the Yangtze River Economic Belt.